基于神经网络的多变量非线性自适应解耦控制研究

提出神经网络前馈自适应解耦控制算法，该算法将多变量非线性系统在平衡点处利用Taylor公式展开，分为线性部分和高阶非线性部分。这样，将高阶非线性部分的影响视为可测干扰，采用前馈补偿的方法加以消除，就可以借助多变量线性系统的自适应解耦控制算法，实现多变量非线性系统的自适应解耦控制，这种方法可以取消被解耦系统为最小相位的限制。
关键词：多变量非线性 解耦控制 自适应控制 神经网络控制
Abstract: The paper put forward an algorithm of NN feedback adaptive decoupling control, and the algorithm outspread the multivariable system with the Taylor formula in the point of balance, which contains linear part and high rank nonlinear part. Thus we may regard the impact of high rank nonlinear part as measurable disturb, then we can call it off with feedback compensation and realize adaptive decoupling control to the multivariable nonlinear system with the help of the multivariable linear adaptive decoupling control algorithm. This method can call off the restrict of the least phase for the decoupled system.
Keywords: multivariable; nonlinear; decoupling control; adaptive control; NN control